[clue-tech] MySQL server reply packets delayed/lost under network congestion?

[Jim Ockers]

Hi CLUEbies,

We have a Windows MySQL client on one end of a 1Mbps serial PPP 
connection and a Linux MySQL server on the other end.  Actually we have 
thousands of these and they all behave the same.  Since it's weird I'd 
appreciate any ideas or suggestions about what we could change or do to 
fix this!

Whenever there is sustained high bandwidth network traffic over the PPP 
link (such as downloading a large e-mail attachment, or streaming video) 
the mysql connections all start to time out because response packets 
from the server are not received at the client.  I observed this by 
doing a simple "telnet server 3306" and observed that the MySQL banner 
response with version number was delayed by several seconds or until the 
bandwidth-hogging stopped.

What I don't understand is why I can still transfer data from other 
servers on the Linux system, such as "net view \\server" and "dir 
\\server\sharename" (using the Windows redirector to talk to Samba on 
the Linux system).  Also the apache web server on the Linux system 
responds normally, both from a web browser and also in response to 
"telnet server 80" and then "GET / HTTP/1.0".

It seems that every network service on the linux server responds 
normally and in a timely manner in the presence of a bandwidth hog, 
EXCEPT for MySQL.  The link MTU is 1500 bytes on both ends, and server 
interface txqueuelen is 100.  MySQL seems to be somewhat better at 
responding when we decrease the MTU to 384, but for a variety of reasons 
that is not practical and we are going to have to stick with 1500.

Is this behavior expected?  Has anyone else seen this?  Is there 
something we can change in the mysql server config which will make it 
more aggressive about sending query response packets out to clients?  By 
the way both 4.0.24 and 5.0.45 do this, although it seems the older 
version is slightly more robust.  I am not a MySQL expert so I don't 
know what the differences are or why this would happen.

We can use tc traffic shaping to force it to work better, but traffic 
shaping has its own set of issues and risks, and really the only thing 
on the server that's not working OK in the presence of a bandwidth hog 
is the mysql server.  Any ideas, anyone?

Thanks,
Jim

-- 
Jim Ockers, P.Eng. (ockers at ockers.net)
Contact info: http://www.ockers.ca/pason.html


-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://cluedenver.org/pipermail/clue-tech/attachments/20091106/9339a545/attachment.html